TT-WMD-2, APPENDIX II
Appendix II
WIGOS Metadata Standard
WIGOS Metadata – Semantic Standard
ICG-WIGOS TT-WMD
CIMO: Brian Howe, Environment Canada, Canada (Chair)
CBS: Karl Monnik, Bureau of Meteorology, Australia
JCOMM: Joe Swaykos, NOAA National Data Buoy Center, United States
CCl: Manuel Bañón Garcia, Antonio Mestre, State Meteorological Agency (AEMET), Spain
CAeM: Stewart Taylor, Met Office, United Kingdom
Member: ZHAO Licheng, China Meteorological Administration, China
CHy: Tony Boston, Bureau of Meteorology, Australia
CAS: JörgKlausen, Federal Office of Meteorology and Climatology MeteoSwiss, Switzerland
Associate Member: Tim Oakley (GCOS)
WMO
Roger Atkinson, Steve Foreman, Luis Nunes
Version 0.1.01
15 May 2014
Version Control
Version / Date / Who / What0.0.0 / 2013-06-06 / J. Klausen / Consolidate input received from Brian Howe after TT-WMD telecom-2
0.0 / 2013-06-06 / J. Klausen / Same as v0.0.0 w/o track changes; new definition of 1-04, code list 1-05
0.0.1 / 2013-06-10 / J. Klausen / Included content for category 4 (environment)
0.0.2 / 2013-06-30 / S Taylor / Included content for category 10 (contact)
0.0.3 / 2013-07-01 / T Boston / Edits to category 7 (station/platform)
0.0.4 / 2013-07-02 / K Monnik
0.0.5 / 2013-07-16 / J. Klausen, B. Howe / Version after Telecon-3
0.0.6 / 2013-07-18 / T. Boston / Edits to category 4 (environment), category 7 (station/platform); code tables 4-02; 7-03
0.0.7 / 2013-08-06 / J. Klausen / After Telecon-4
0.0.8 / 2013-09-0208-29 / T. Boston, B. Howe / Edits to topographycategory 5 and platform/station modelcorresponding code table.
0.0.9 / 2013-09-03 / J. Klausen / After Telecon-5
0.0.10 / ?? / ?? / Intermediate version of uncertain origin
0.0.11 / 2013-10-3 / J,.Klausen / After Telecon-6, with expansions not discussed during telecom
0.0.12 / 2013-10-03 / B. Howe / After Telecon-6 with changes accepted.
0.0.13 / 2013-10-24 / B. Howe / After Telecon-7
0.0.13.ra / 2013-10-31 / R. Atkinson / Responses to a number of comments in 0.0.13
0.0.13.ra+km / 2013-11-04 / K. Monnik / General edits, additions to Cat 8, added examples to Cat 1, 5, 7.
0.0.14 / 2013-11-04 / J. Klausen / After Telecon-8
0.0.14 km / 2013-11-06 / K. Monnik / Minor changes to 6.06, 8.03, 8.10, plus selected comments from Blair Trewin (AU)
0.0.15 / 2013-11-11 / J. Klausen / After Telecon-9, and including feed-back from P. Pilon/R. Atkinson
0.0.16 / After Telecon-10
0.0.17 / 2013-12-19 / J. Klausen / After Telecon-11
0.0.18 / 2014-02-06 / J. Klausen, K Monnik / Response to WielWauben, Bruce Forgan; version after Telecon-12, with further additions and edits, formatting
0.0.19
0.0.20 / 2014-03-12
2014-03-18 / B. Howe
J. Klausen / After Telecon-15, accepted ICG-WIGOS MCO classifications and added two requested fields. Numerous other updates accepted.
Comments by ET-SUP carried over.
0.0.21 / 2014-03-27 / J. Klausen / Element 5-04 (Reporting interval (space)) explicitly listed; code table 5-05 included; element 5-11 (reference time) defined and explained; numbering in list of category 5 corrected; Figures 1 and 2 updated
0.0.22 / 2014-04-03 / J. Klausen / After Telecon-16
0.0.23 / 2014-04-28 / J. Klausen / After Telecon-17, several changes accepted, minor editing, fixed a few cross-references
0.1 / 2014-05-15 / J. Klausen / Version after TT-WMD-2; dropped notion of “Core” in favor of a phased implementation; added element 8-00; dropped 4-04; moved element 8-05 to become 4-04; editorial improvements
0.1.01 / 2014-05-19 / WIGOS-PO / Editorial
Table of Contents
WIGOS Metadata Standard
Purpose and Scope of WIGOS Metadata
WIGOS Metadata Categories
A Note on Space and Time
Reporting Obligations for WIGOS Metadata
Implementation and Use of Standard
Adoption through a Phased Approach
Category 1: Observed Quantity
Category 2: Purpose of Observation
Category 3: Data Quality
Category 4: Environment
Category 5: Data Processing and Reporting
Category 6: Sampling and Analysis
Category 7: Station/Platform
Category 8: Method of Observation
Category 9: Ownership & Data Policy
Category 10: Contact
References
Instructions for Developers
Purpose and Scope of WIGOS Metadata
An important aspect of WIGOS (WMO Integrated Global Observing System) implementation is ensuring maximum usefulness of WIGOS observations and measurement data. Data on its own is of very limited use: it is only when accompanied by adequate metadata (data describing the data) that the full potential of the data can be utilized. Metadata of two complementary types are required. The first of these is discovery metadata – information that facilitates data discovery, access and retrieval. These metadata are WIS (WMO Information System) metadata and are specified and handled as part of WIS. The second type is interpretation or description metadata – information that enables data values to be interpreted in context. These latter metadata are WIGOS metadata and are the subject of this specification, which provides a WIGOS-wide standard for the interpretation metadata required for the effective interpretation of data from all WIGOS observing sub-systems by all data users.
WIGOS metadata should describe the observed quantity, the conditions under which it was observed, how it was measured, and how the data has been processed, in order to provide data users with confidence that the use of the data is appropriate for their application. GCOS Climate Monitoring Principle #3 describes the relevance of metadata as:
“The details and history of local conditions, instruments, operating procedures, data processing algorithms and other factors pertinent to interpreting data (i.e., metadata) should be documented and treated with the same care as the data themselves.”
WIGOS Observations consist of an exceedingly wide range of data from the manual observations to complex combinations of satellite hyper-spectral frequency bands, measured in situ or remotely, from single dimension to multiple dimensions, and those involving post observation analysis. A comprehensive metadata specification to cover all types of data is by nature complex to define. A user should be able to use the WIGOS metadata to identify the conditions under which the observations or measurement was made, and any aspects which may affect the use or understanding of the data; i.e. to determine whether the data are fit for purpose.
WIGOS Metadata Categories
Ten categories of metadata have been identified. These are listed in Table 1below.They define the WIGOS metadata standard. All of the elements listed are considered to be important for the documentation and interpretation of observations made even in the distant future. Hence, the standard currently declares many elements as mandatory that are clearly not needed for applications focusing on more immediate use of observations. For these applications, such as numerical weather prediction, aeronautical or other transport sector applications, advisories, etc., profiles of the standard would have to be developed. The categories are in no particular order but reflect the need to specify the observed quantity; to answer why, where, and how an observation was made; how the raw data were processed; and what the quality of the data is.
Each of these categories contains a number of individual elements as shown in Figure 1. Note that some of these elements will most likely be implemented using several individual entities (e.g., geolocation will consist of the atomic elements latitude, longitude, elevation or a set of polar coordinates.)
Table 1. WIGOS metadata categories
# / Category / Description1 / observed quantity / Groups elements that specify the observed quantity and the data sets generated. It includes an element describing the spatial representativeness of the observations as well as the biogeophysical compartment the observations describe.
2 / purpose of observation / Specifies the main application area(s) of an observation and the observation program(s) an observation is affiliated to.
3 / data quality / Specifies the data quality and traceability of an observation or dataset.
4 / environment / Describes the geographical environment within which the observations are made. It also provides an unstructured element for additional meta-information that is considered relevant for adequate use of the data and that is not captured anywhere else in the standard.
5 / data processing and reporting / Specifies how raw data are transferred into the reported physical quantities.
6 / sampling and analysis / Specifies how sampling and/or analysis are used to derive the reported observation or how a specimen was collected
7 / station/platform / Specifies the environmental monitoring facility, including fixed station, moving equipment or remote sensing platform, at which an observation is made.
8 / method of observation / Specifies the method of observation and describes characteristics of the instrument(s) used to make the observation. If multiple instruments need to operate in concert to arrive at an observation, then this category should be repeated.
9 / ownership and data policy / Specifies who is responsible for the observation and owns it.
10 / contact / Specifies where information about an observation or dataset can be obtained.
For example, an observation / dataset will have the following metadata categories associated with it
•One or several purpose(s) of observation (e.g. upper air observations and surface synoptic observations)
•Data processing procedures associated with the instruments
•Instruments which have been used to make the observation
•A station/platform to which the instrument(s) belong(s)
•Ownership and data policy restriction
•Contact
An instrument can observe/measure one or more quantities. For example:
•a resistance temperature device can report temperature;
•a humidity probe can report temperature and humidity;
•a sonic anemometer can report wind speed, wind direction and air temperature
An instrument can be associated with:
•sampling and analysis (e.g. 10 Hz samples of air temperature)
•data processing (e.g. ceilometer reporting of 10 min statistics of cloud height following processing through sky condition algorithm);
An observed quantity can be influenced or characterized by the environment, for example:
•wind speed (observed quantity) on top of a hill (environment);
•river yield (observed quantity) characterized by the upstream catchment and landuse
Page 1 of 45
TT-WMD-2, APPENDIX II
Figure 1. UML diagram specifying the WIGOS Metadata Standard (**: code tables expected; [0..1*]: optional or conditional elements. Conditional elements become mandatory if a given condition is met. Conditions are referenced in parentheses. Optional elements may be declared mandatory as part of profiling the standard for specific application areas; [1..*]: mandatory elements. These elements must be reported, and if no value is available, a nilReason must be specified)
Page 1 of 45
TT-WMD-2, APPENDIX II
A Note on Space and Time
It is important to understand that WIGOS metadata are intended to describe a dataset, i.e. one or several observations, including the where, when, how, and even why the observations were made. As a consequence, reference to space and time is made in several places throughout the standard.
Figure 2 illustrates the concepts and terms used to describe the temporal aspects of an observation or dataset, sampling strategy, analysis and data processing.
The concepts and terms used to describe spatial aspects (i.e., geolocation) of observations are perhaps even more complex (cf. ). For example, for ground-based in-situ observations, the spatial extent of the observation coincides with the geolocation of the sensor, which in most cases will be time-invariant and is normally close to the geolocation of the station/platform where the observation was made. For a satellite-based lidar system, the situation is quite different. Depending on the granularity of metadata desired, the spatial extent of the individual observation may be an individual pixel in space, the straight line probed during an individual laser pulse, or perhaps an entire swath. In any case, the spatial extent of the observation will not coincide with the location of the sensor. The WIGOS metadata standard therefore needs to take into account such quantities as
- the spatial extent of the observed quantity (e.g. atmospheric column above a Dobson Spectrophotometer) (cf. 1-04)
- the location of the station/platform (e.g. radar transmitter/receiver or aircraft position/route) (cf. 7-07)
- the location of the instrument (e.g. the anemometer is located adjacent to Runway 23) (cf. 8-04, 8-11)
- the spatial representativeness of the observation (cf. 1-05)
All these are expressed in terms of geolocations, specifying either a zero-dimensional geographic extent (a point), a one-dimensional geographic extent (a line, either straight or curved), a two-dimensional geographic extent (a plane or other surface), or a three-dimensional geographic extent (a volume).
By way of example, a station/platform can be:
- collocated with the observed quantity as for in situ surface observations station (e.g. AWS)
- collocated with the instrument but remote to the observed quantity (e.g. Radar)
- remote from where the instrument may transmit data to the station (e.g. Airport surface station where instruments are located across the airport, or a balloon atmosphere profiling station)
- in motion and travelling through the observed medium (e.g. Aircraft AMDAR equipped aircraft)
- in motion and remote to the observed medium (e.g. satellite platform)
An Instrument can be:
- collocated with the observed quantity (e.g. surface observations temperature sensor);
- remote to the observed quantity (e.g. radar transmitter/receiver);
- in motion but located in the observed medium (e.g. radiosonde)
- in motion and remote from the observed quantity (e.g. satellite based radiometer)
- located within a standardized enclosure (e.g. a temperature sensor within a Stevenson screen)
Figure 2. Graphical representation of temporal elements referenced in WIGOS Metadata categories
Figure 3. Graphical representation of spatial elements referenced in WIGOS Metadata categories
Reporting Obligations for WIGOS Metadata
In keeping with ISO, the elements are classified as either mandatory (M), conditional (C), or optional (O).
Most of the elements in this standard are considered mandatory and presumed to be of importance for all WMO Technical Commissions in view of enabling adequate future use of (archived) data. Metadata providers are expected to report mandatory metadata elements, and a formal validation of a metadata record will fail if mandatory elements are not reported. The TT-WMD recognizes however that not all Members may be in a position to provide all these elements at present, and that a small fraction of these mandatory elements may not even be applicable for a specific type of observation. Under such circumstances, the reason for not providing information on mandatory elements shall be reported as “not applicable” or “unknown”. The motivation for this is that knowledge of the reason why a mandatory metadata element is not available provides more information than not reporting a mandatory element at all. In the tables below, these cases are indicated with M#.
Optional elements provide useful information that can help to better understand an observation. In this version of the standard, very few elements are considered optional. Optional elements are likely to be important for a particular community, but less so for others.
Conditional elements must be reported under certain conditions. For example, the element “Reporting interval (space) is classified as conditional, because it only applies to remote sensing observations andmobile platforms. Therefore, the elements in this category should be considered mandatory for remote sensing andmobile observing systems but not so for e.g., surface land stations.
Implementation and Use of Standard
This document is a semantic standard, not an implementation guideline. A semantic standard specifies the elements that exist and that can be reported.It does not specify how the information shall be encoded or exchanged. However, the following are likely scenarios and important aspects that may help the reader appreciate what lies ahead.
- The most likely implementation will be in XML, in line with the specifications for WIS metadata and common interoperability standards. Regardless of the final implementation, the full metadata record describing a dataset can be envisioned as a tree with the category as branches off the stem, and the individual elements as leaves on these branches. Some branches may occur more than once, e.g., a dataset may have been generated using more than one instrument at once, in which case two branches for ‘instrument’ may be required.
- Not all of the elements specified in this document need to be updated at the same frequency. Some elements, such as position of a land-based station are more or less time-invariant, while others, such as a specific sensor, may change regularly every year. Still other elements, such as environment, may change gradually or rarely, but perhaps abruptly.Finally, elements restricting the application of an observation, e.g., to road condition forecasting, may have to be transmitted with every observation. The implementation of the WIGOS metadata needs to be able to deal with this.
- Not all applications of observations require the full suite of metadata as specified in this standard at any given time. The amount of metadata that needs to be provided to be able to make adequate use of an observation, for example for the purpose of issuing a heavy precipitation warning, is much less than for the adequate use of even the same observation for a climatological analysis. On the other hand, the metadata needed for near-real-time applications also need to be provided in near-real-time. This is important to realize, as it makes the task of providing WIGOS metadata much more tractable. The implementation of WIGOS metadata needs to be able to cope with vastly different update intervals, and incremental submission of additional metadata to allow the creation of ‘complete’ metadata records.
- User will want to obtain and filter datasets according to certain criteria / properties as described within each WIGOS metadata record. This functionality requires either a central repository for WIGOS metadata or full interoperability of the archives collecting WIGOS metadata.
How, then can these requirements be met? In the case where observations are clearly only used for some near-real-time application and there is clearly no long-term use or re-analysis application to be expected, a profile of the WIGOS metadata standard may be specified that declares a specific subset of metadata elements as mandatory. This is depicted schematically in Figure 4.
Figure 4. Schematic of the relationship of WIS and WIGOS metadata and the scope of the ISO19115 standard. The WMO Core is a profile of ISO19115. There is clearly an overlap between WIS and WIGOS metadata, but it is undecided if WIS metadata should be viewed as a subset of WIGOS metadata or if certain WIS metadata elements will not be included in WIGOS. WIGOS metadata exceed the scope of the ISO19115 standard. Also shown is a possible profile (subset) of WIGOS metadata elements for some specific near-real-time application.
Importantly, all WIGOS metadata elements (or group of elements) will have to be time-stamped with the time of validity and associated to a unique identifier for a dataset during transmission and for archiving. Using this approach, increments of a ‘full’ WIGOS metadata record can be transmitted anytime changes occur and updates are deemed necessary. At the archive, the increments can be added to the existing metadata record for that dataset, in time establishing the full history of a particular observation.